DeepSeek 4.1

Large language model from DeepSeek, version 4.1, used via chat and API.

Commenters use it as a daily driver, including for frontend and backend code,123 and like it as a cheap model that hallucinates less than 4.0 and costs far less than a Claude subscription.45 The main caveat is that one daily user calls it nowhere close to top models,6 and it ranks low on at least one benchmark.7
The HN verdict, written from 7 cited comments
5recommend
0warn against
Sep 12, 2026first mention
Oct 26, 2026latest mention
Nov 25 – Oct 26
Used forfor codingbudget-friendlycheap modeldaily driver model

Why people recommend it

Daily driver · 3

Several commenters use it as their main everyday model.

“DSv4.1 is my daily driver.”

“i've transitioned to using DS 4.1 as a daily driver”

Show 1 more

Low cost · 3

Valued as a cheap model compared with paid Claude subscriptions and similar models.

“I'm glad I have 4.1 now for that.”

Show 1 more

“Deepseek 1017 in 6.5 minutes”

Coding capability · 2

Commenters say it handles real coding work well, and one prefers it to GPT6 Sol.

“i've transitioned to using DS 4.1 as a daily driver”

“Not even GPT6 Sol cannot match DeepSeek 4.1 in my work”

Fewer hallucinations · 1

One commenter says it hallucinates less than DeepSeek 4.0.

“I'm glad I have 4.1 now for that.”

Caveats and dissent

Not top-tier · 1

One daily user says it is nowhere close to leading models.

“I use Deepseek 4.1 almost every day as well, it's nowhere close”

Benchmark ranking · 1

It ranks very low on at least one benchmark, though the commenter still finds it capable.

“i've transitioned to using DS 4.1 as a daily driver”

Speed and tokens · 1

One commenter says Chinese models generally use more tokens and finish more slowly than Sonnet.

“Deepseek 1017 in 6.5 minutes”

Often compared with

All comments 7

“Deepseek 1017 in 6.5 minutes”

“I'm glad I have 4.1 now for that.”

“DSv4.1 is my daily driver.”

“i've transitioned to using DS 4.1 as a daily driver”

“I use Deepseek 4.1 almost every day as well, it's nowhere close”

“Not even GPT6 Sol cannot match DeepSeek 4.1 in my work”

Threads where it came up

Also written as DeepSeek V4.1. Data: deepseek-4-1.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.

Keyboard shortcuts
j / kNext / previous item
o or EnterOpen selected item
/Search
g hHome
g tTop
g rRising
g cCategories
tToggle dark mode
?This help